Bareos high availability

[Chanaka Madushan]

Hi,

I got a requirement for install bareos as a high available cluster on CentOS with PostgreSQL database. But in this case I have limited to two servers + a shared storage (may be a SAN or a NAS). So I have to install all bareos-fd, bareos-sd and bareos-dir on these two servers as high available services.

I hope to make PostgreSQL high availability with transaction log shipping.

But I donot have an idea to make bareos services high available.

Is there anyone who has deployed a high available cluster for bareos?

[John Goutbeck]

Hello;

I to have this requirement to have Bareos 19.2 HA on my cluster.  Does Anybody have a write up on this subject?

My cluster is SLES 15 SP1 + HA, DRBD storage

I have 1 small DRBD resourse (100GB) for the postgresql and another DRBD resource (13TB) for the Bareos storage

The idea is to have the DRBD storage, Postgresql & Bareos migrate/failover to another node.

Have a good day.

[Oleg Volkov]

I do not see any problem. It is just a service.

Make postgres HA, make /etc/bareos and  /var/lib/bareos be on shared disk, make VIP and colocate it with bareos services.

Never tried this with bareos, but made a lot of clusters - there should be no problem.

Just follow any active-standby scenario.

[Spadajspadaj]

I would be, however, cautious about possible scenarios where a node breaks and fails over to the other server - for example - in the middle of a backup job. Such scenarios would need some testing so you know what to expect and how to handle such situation.

--
You received this message because you are subscribed to the Google Groups "bareos-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to bareos-users...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/bareos-users/3833dc30-dac4-47d2-8c00-3d13c0fbe6c6o%40googlegroups.com.

[Oleg Volkov]

Jobs will be obviously aborted and failed.

Then you have to care about them manually as usual for failed job.

K.O.

On Friday, June 5, 2020 at 1:32:40 PM UTC+3, Spadajspadaj wrote:

I would be, however, cautious about possible scenarios where a node breaks and fails over to the other server - for example - in the middle of a backup job. Such scenarios would need some testing so you know what to expect and how to handle such situation.

On 05.06.2020 09:00, Oleg Volkov wrote:

I do not see any problem. It is just a service.

Make postgres HA, make /etc/bareos and  /var/lib/bareos be on shared disk, make VIP and colocate it with bareos services.

Never tried this with bareos, but made a lot of clusters - there should be no problem.

Just follow any active-standby scenario.

On Tuesday, June 20, 2017 at 8:47:27 AM UTC+3, Chanaka Madushan wrote:

Hi,

I got a requirement for install bareos as a high available cluster on CentOS with PostgreSQL database. But in this case I have limited to two servers + a shared storage (may be a SAN or a NAS). So I have to install all bareos-fd, bareos-sd and bareos-dir on these two servers as high available services.

I hope to make PostgreSQL high availability with transaction log shipping.

But I donot have an idea to make bareos services high available.

Is there anyone who has deployed a high available cluster for bareos?

--
You received this message because you are subscribed to the Google Groups "bareos-users" group.

To unsubscribe from this group and stop receiving emails from it, send an email to bareos...@googlegroups.com.

[Spadajspadaj]

Probably. That's why I'd probably prefer some duplicated, but not necessarily HA setup. I'm always a bit suspicious towards HA-clustering non-HA-capable solutions.

But YMMV

To unsubscribe from this group and stop receiving emails from it, send an email to bareos-users...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/bareos-users/fead0ad8-344d-4a1f-8eb1-84affc638961o%40googlegroups.com.

[John Goutbeck]

Hello Oleg;

Thank you for this.  It should be this simple to HA a solution. 

May time we tend to overthink the issue and complicate the solution. 

Yes the will be issues upon a failover, but that is to be expected.  Just deal with the issue and carry on.

Have a good day.

--

You received this message because you are subscribed to the Google Groups "bareos-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to bareos-users...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/bareos-users/3833dc30-dac4-47d2-8c00-3d13c0fbe6c6o%40googlegroups.com.

[John Goutbeck]

So I can HA the postgresql (on a drbd volume) ( postgresql block device) on my SLES 15 HA KVM cluster.  Postgresql has a ocf:heartbeat resource definition file.

Soft link the /etc/bareos and the /etc/bareos-webui folders to the postgresql drbd block device, so they follow the drbd primary resource on either SLES HA host.

Now on to HA Bareos.  I do not see a ocf:heartbeat that I can use to stop/start the 3 bareos service.  Can anybody offer any help on how to HA resource the bareos dir/fd/sd?

Have a good day

[Oleg Volkov]

Once you on SLES15, this command could help you to find all bareos systemd services on server:

# crm configure ra list systemd
Also, you can use HAWK and configure it using wizards. Look for services in "systemd" section.

To unsubscribe from this group and stop receiving emails from it, send an email to bareos...@googlegroups.com.